smagt/meditron-7b-instruct
TEXT GENERATIONConcurrent Unit Cost:1Model Size:7BQuant:FP8Context Size:4kPublished:May 19, 2024License:llama2Architecture:Transformer Open Weights Featherless Exclusive Cold
smagt/meditron-7b-instruct is a 7 billion parameter instruction-tuned Llama-based model developed by Nicolai van der Smagt, building upon epfl-llm/meditron-7b. Fine-tuned for 5 epochs on the Databricks Dolly-15k dataset, it specializes in instruction-following tasks. With a 4096-token context length, this model is designed for general-purpose conversational AI, particularly in English.
Loading preview...
Meditron 7B Instruct: Instruction-Tuned Llama Model
smagt/meditron-7b-instruct is a 7 billion parameter language model developed by Nicolai van der Smagt, based on the Llama architecture. It is an instruction-tuned variant of the epfl-llm/meditron-7b base model, specifically fine-tuned for 5 epochs using the databricks/databricks-dolly-15k dataset.
Key Capabilities
- Instruction Following: Optimized to understand and execute instructions provided in natural language, making it suitable for a wide range of conversational and task-oriented applications.
- General Purpose: While its base model has medical associations, this instruction-tuned version, trained on Dolly-15k, is adapted for broader instruction-following tasks.
- English Language Support: Primarily designed for English language processing.
- Context Length: Supports a context window of 4096 tokens, allowing for processing moderately long inputs.
Good For
- Conversational AI: Developing chatbots or virtual assistants that respond to user prompts and instructions.
- Task Automation: Automating simple text-based tasks through instruction-based interactions.
- Research and Development: Experimenting with instruction-tuned Llama models on a consumer-grade GPU setup due to its 7B parameter size.